Introduction

In this project, we are going to use spacy for entity recognition on 200 Resume and experiment around various NLP tools for text analysis. The main purpose of this project is to help recruiters go throwing hundreds of applications within a few minutes. We have also added skills match feature so that hiring managers can follow a metric that will help them to decide whether they should move to the interview stage or not. We will be using two datasets; the first contains resume texts and the second contains skills that we will use to create an entity ruler.

Dataset

livecareer.com resume Dataset

A collection of 2400+ Resume Examples taken from livecareer.com for categorizing a given resume into any of the labels defined in the dataset: Resume Dataset.

Import

#spacy
import spacy
from spacy.pipeline import EntityRuler
from spacy.lang.en import English
from spacy.tokens import Doc

#gensim
import gensim
from gensim import corpora

#Visualization
from spacy import displacy
import pyLDAvis.gensim_models
from wordcloud import WordCloud
import plotly.express as px
import matplotlib.pyplot as plt

#Data loading/ Data manipulation
import pandas as pd
import numpy as np
import jsonlines

#nltk
import re
import nltk
from nltk.corpus import stopwords
from nltk.stem import WordNetLemmatizer
nltk.download(['stopwords','wordnet'])

#warning
import warnings 
warnings.filterwarnings('ignore')

[nltk_data] Downloading package stopwords to
[nltk_data]     /home/dristanta/nltk_data...
[nltk_data]   Package stopwords is already up-to-date!
[nltk_data] Downloading package wordnet to
[nltk_data]     /home/dristanta/nltk_data...
[nltk_data]   Package wordnet is already up-to-date!

## For dropdown
import numpy as np
import pandas as pd
import textwrap

import ipywidgets as widgets
from ipywidgets import interact, interact_manual
import IPython.display
from IPython.display import display, clear_output

import plotly.graph_objects as go

Resume Dataset

Using Pandas read_csv to read dataset containing text data about Resume.

• we are going to randomized Job categories so that 200 samples contain various job categories instead of one.
• we are going to limit our number of samples to 1000 as processing 2400+ takes time.

df = pd.read_csv("Resume.csv")

df = df.reindex(np.random.permutation(df.index))
data = df.copy().iloc[0:1000,]
data.head()

	ID	Resume_str	Resume_html	Category
1545	23573064	PROGRAMME FINANCE ASSOCIATE Pro...	<div class="fontsize fontface vmargins hmargin...	FINANCE
2310	17033567	VIDEO DIRECTOR, EAST COAST VIDEO FOR ...	<div class="fontsize fontface vmargins hmargin...	ARTS
794	10268614	FITNESS ATTENDANT Summary ...	<div class="fontsize fontface vmargins hmargin...	FITNESS
308	40018190	IT SUPPORT TECHNICIAN Education...	<div class="fontsize fontface vmargins hmargin...	INFORMATION-TECHNOLOGY
770	29134372	REGIONAL RECRUITER Summary M...	<div class="fontsize fontface vmargins hmargin...	HEALTHCARE

nlp = spacy.load("en_core_web_lg")

skill_pattern_path = "jz_skill_patterns.jsonl"

Entity Ruler

To create an entity ruler we need to add a pipeline and then load the .jsonl file containing skills into ruler. As you can see we have successfully added a new pipeline entity_ruler. Entity ruler helps us add additional rules to highlight various categories within the text, such as skills and job description in our case.

ruler = nlp.add_pipe("entity_ruler")
ruler.from_disk(skill_pattern_path)
nlp.pipe_names

['tok2vec',
 'tagger',
 'parser',
 'attribute_ruler',
 'lemmatizer',
 'ner',
 'entity_ruler']

Skills

We will create two python functions to extract all the skills within a resume and create an array containing all the skills. Later we are going to apply this function to our dataset and create a new feature called skill. This will help us visualize trends and patterns within the dataset.

• get_skills is going to extract skills from a single text.
• unique_skills will remove duplicates.

def get_skills(text):
    doc = nlp(text)
    myset = []
    subset = []
    for ent in doc.ents:
        if ent.label_ == "SKILL":
            subset.append(ent.text)
    myset.append(subset)
    return subset


def unique_skills(x):
    return list(set(x))

Cleaning Resume Text

We are going to use ne ltk library to clean our dataset in a few stepse e :

• We are going to use regex to remove hyperlinks, special characters, or punctuations.
• Lowering text
• Splitting text into array based on space
• Lemmatizing text to its base form for normalizations
• Removing English stopwords
• Appending the results into an array.

clean = []
for i in range(data.shape[0]):
    review = re.sub(
        '(@[A-Za-z0-9]+)|([^0-9A-Za-z \t])|(\w+:\/\/\S+)|^rt|http.+?"',
        " ",
        data["Resume_str"].iloc[i],
    )
    review = review.lower()
    review = review.split()
    lm = WordNetLemmatizer()
    review = [
        lm.lemmatize(word)
        for word in review
        if not word in set(stopwords.words("english"))
    ]
    review = " ".join(review)
    clean.append(review)

Applying functions

In this section, we are going to apply all the functions we have created previously

• creating Clean_Resume columns and adding cleaning Resume data.
• creating skills columns, lowering text, and applying the get_skills function.
• removing duplicates from skills columns.

data["Clean_Resume"] = clean
data["skills"] = data["Clean_Resume"].str.lower().apply(get_skills)
data["skills"] = data["skills"].apply(unique_skills)
data.head()

	ID	Resume_str	Resume_html	Category	Clean_Resume	skills
1545	23573064	PROGRAMME FINANCE ASSOCIATE Pro...	<div class="fontsize fontface vmargins hmargin...	FINANCE	programme finance associate professional summa...	[nim, security, support, project management, d...
2310	17033567	VIDEO DIRECTOR, EAST COAST VIDEO FOR ...	<div class="fontsize fontface vmargins hmargin...	ARTS	video director east coast video enterprise bra...	[business]
794	10268614	FITNESS ATTENDANT Summary ...	<div class="fontsize fontface vmargins hmargin...	FITNESS	fitness attendant summary highly motivated nut...	[engineering, schedule, finance, marketing, in...
308	40018190	IT SUPPORT TECHNICIAN Education...	<div class="fontsize fontface vmargins hmargin...	INFORMATION-TECHNOLOGY	support technician education bachelor science ...	[telephony, support, design, project managemen...
770	29134372	REGIONAL RECRUITER Summary M...	<div class="fontsize fontface vmargins hmargin...	HEALTHCARE	regional recruiter summary motivated program m...	[security, support, documentation, accounting,...

Visualization

fig = px.histogram(
    data, x="Category", title="Distribution of Jobs Categories"
).update_xaxes(categoryorder="total descending")
fig.show()

Job_cat = data["Category"].unique()
Job_cat = np.append(Job_cat, "ALL")

Job_cat

array(['FINANCE', 'ARTS', 'FITNESS', 'INFORMATION-TECHNOLOGY',
       'HEALTHCARE', 'CONSTRUCTION', 'CONSULTANT', 'DESIGNER', 'BANKING',
       'TEACHER', 'CHEF', 'AVIATION', 'ADVOCATE', 'BPO',
       'BUSINESS-DEVELOPMENT', 'PUBLIC-RELATIONS', 'AUTOMOBILE',
       'ENGINEERING', 'APPAREL', 'AGRICULTURE', 'DIGITAL-MEDIA',
       'ACCOUNTANT', 'SALES', 'HR', 'ALL'], dtype=object)

def dropdown_menu_widget(Job_Category):
    output=widgets.Output()
    drop_Job=widgets.Dropdown(options = Job_Category, value=None, description='Job Category:')
    
    for cat in Job_Category:
        def dropdown_Job_eventhandler(change):
            display(input_widgets)
            job_choice=change.new
            IPython.display.clear_output(wait=True)
    drop_Job.observe(dropdown_Job_eventhandler,names='value')
    input_widgets=widgets.HBox([drop_Job])
    display(input_widgets)
    IPython.display.clear_output(wait=True)

dropdown_menu_widget(Job_cat)

Job_Category="INFORMATION-TECHNOLOGY"

Total_skills = []
if Job_Category != "ALL":
    fltr = data[data["Category"] == Job_Category]["skills"]
    for x in fltr:
        for i in x:
            Total_skills.append(i)
else:
    fltr = data["skills"]
    for x in fltr:
        for i in x:
            Total_skills.append(i)

fig = px.histogram(
    x=Total_skills,
    labels={"x": "Skills"},
    title=f"{Job_Category} Distribution of Skills",
).update_xaxes(categoryorder="total descending")
fig.show()

Most used words

In this part, we are going to display the most used words in the Resume filter by job category. In Information technology, the most words used are system, network, and database. We can also discover more patterns by exploring the word cloud below.

text = ""
for i in data[data["Category"] == Job_Category]["Clean_Resume"].values:
    text += i + " "

plt.figure(figsize=(8, 8))

x, y = np.ogrid[:300, :300]

mask = (x - 150) ** 2 + (y - 150) ** 2 > 130 ** 2
mask = 255 * mask.astype(int)

wc = WordCloud(
    width=800,
    height=800,
    background_color="white",
    min_font_size=6,
    repeat=True,
    mask=mask,
)
wc.generate(text)

plt.axis("off")
plt.imshow(wc, interpolation="bilinear")
plt.title(f"Most Used Words in {Job_Category} Resume", fontsize=20)

Text(0.5, 1.0, 'Most Used Words in INFORMATION-TECHNOLOGY Resume')

Entity Recognition

We can also display various entities within our raw text by using spaCy displacy.render. I am in love with this function as it is an amazing way to look at your entire document and discover SKILL or GEP within your Resume.

sent = nlp(data["Resume_str"].iloc[0])
displacy.render(sent, style="ent", jupyter=True)

PROGRAMME FINANCE SKILL ASSOCIATE Professional Summary Seeking a position where I can demonstrate my skills and contribute to an organization that offers professional growth. Being productive and add value to the organisation through my knowledge and previous experiences Overview Certified project manager with two years DATE experience in managing projects according to PRINCE2 ORG methodology. Civil engineering SKILL background. Two years DATE experience in events organization (workshops, conferences and awareness campaigns). Highly motivated in team and individual tasks, hard working and productive worker under pressure. Practical experience of dealing with governmental and official entities. Outstanding communication skills with fluency in three CARDINAL languages SKILL . Good knowledge of international and development practices. Management and administrative professional. People oriented with strong leadership abilities. Excellent negotiation skills. Skills Work History 06/2012 CARDINAL to Current Programme Finance SKILL Associate Company Name – City , State Ensured administration and implementation of programme/operations strategies, adapts processes and procedures focusing on achievement of the following results: Full compliance with UN ORG / UNDP ORG rules and regulations and other relevant policies on financial recording/reporting system and follow-up on audit recommendations; Implementation of effective internal controls, proper functioning of a client-oriented financial resources management system. Provision of inputs for implementation of cost-saving and reduction strategies in consultation with the office management. Preparation of cost sharing, trust fund agreements, follow up on contributions within the CO resource mobilization efforts. Ensures proper tracking of donor reporting and ensure with program/projects concerned that reporting is on time. 2 CARDINAL . Provided effective support SKILL to management of the CO programme, administration of budgets and functioning of the optimal cost-recovery system focusing on achievement of the following results: Presentation of information for formulation of country programme work plans, budgets, proposals on implementation arrangements and execution modalities. Entry of data of new grants into Atlas ORG in the form of Annual Work Plans WORK_OF_ART (AWPs), monitoring SKILL of their status. Provision of guidance to the executing agencies on routine implementation of projects, tracking use of financial resources. Accuracy verification of Combined Delivery Reports. Provision of information for the audit of NIM/DIM ORG projects, tracks implementation of audit recommendations. Support SKILL the preparation of the Integrated Work Plan LAW for the country office, under the supervision of the Deputy Country Director. Ensuring that the project tree is properly set up in Atlas ORG , with projects linked to the correct outcomes Tracking of overall office indicators and delivery figures. Tracking and detailed reporting on mobilized resources. Ensure that GMS ORG rates are properly charged to development projects. Provides in puts in the preparation of cost-recovery bills in Atlas PERSON for the services provided by, elaboration and implementation of the income tracking system SKILL and follow up on cost recovery. 3 CARDINAL . Provided accounting SKILL and administrative support SKILL to the Programme Finance Unit ORG focusing on achievement of the following results: Timely corrective actions on erroneous data in Atlas ORG . Processing of GLJE's if required. Presentation of thoroughly researched information for planning of financial resources of The CO, reports containing analysis of the financial situation. 4 CARDINAL . Ensured facilitation of knowledge building and knowledge sharing in the CO focusing on achievement of the following results: Systematic gaining and sharing of knowledge and experience related to programme management. Organization of training for the office staff on programme/operations related issues, including program/project related Atlas ORG training. Synthesis of lessons learned and best practices in programme finance SKILL . Sound contributions to knowledge networks and communities of practice. 10/2008 to CARDINAL 05/2012 Programme Assistant Company Name Tasks and responsibilities 10/2007 CARDINAL to 03/2008 Teacher Company Name Teaching Italian Language as a foreign language. 01/1999 to 01/2002 Office manager and secretary Tasks and responsibilities:. Office management, organization and documentation SKILL . Overseeing financial duties, including payrolls, budgets and spending. Education 12 2008 DATE BSc PRODUCT . Civil Engineering ORG , Al-Fateh University ORG , Tripoli GPE - Libya GPE ( 1997 - 2001 DATE ) & University of Tripoli ORG , Tripoli GPE - Libya GPE ( 2005 - 2006 DATE ) Projects In Controlled Environments (PRINCE2) Foundation, UNDP ORG training, Tripoli GPE - December 2009 DATE Enterprise Risk Management, UNDP ORG training, Tripoli GPE - January 2009 DATE HIV & AID, UNDP ORG training, Tripoli GPE - January 2009 DATE Ethics Train, UNDP ORG training, Tripoli GPE - December 2008 The Gender Journey: Thinking Outside the Box WORK_OF_ART , UNDP ORG training, Tripoli GPE - January 2009 DATE UN Programme On The Prevention Of Harassment, Sexual Harassment and Abuse Of Authority, UNDP ORG training, Tripoli GPE - January 2009 Basic and advanced security SKILL in the field, UNDP ORG training, Tripoli GPE : Skills accounting SKILL , administrative support SKILL , Arabic LANGUAGE , Basic, BSc ORG , budgeting, budgets, budget, consultation, client, Delivery ORG , documentation SKILL , English LANGUAGE , Finance ORG , Financial ORG , grants, interpretation, Italian NORP , Languages SKILL , Director, MS Office ORG , office, outlook, Enterprise ORG , networks, Office ORG management, policies, PRINCE2 ORG , processes, progress, project management SKILL , proposals, quality control, recording, recruitment, reporting, research, Risk Management ORG , Sound, supervision, Teaching, translating, translation Additional Information Date of Birth: 5 CARDINAL November 1978 DATE
Marital status: Married
Nationality: Libyan

Dependency Parsing

We can also visualize dependencies by just changing style to dep as shown below. We have also limited words to 10 which includes space too. Limiting the words will make it visualize the small chunk of data and if you want to see the dependency, you can remove the filter.

displacy.render(sent[0:10], style="dep", jupyter=True, options={"distance": 90})

Custom Entity Recognition

In our case, we have added a new entity called SKILL and is displayed in gray color. I was not impressed by colors and I also wanted to add another entity called Job Description so I started experimenting with various parameters within `displace.

• Adding Job-Category into entity ruler.
• Adding custom colors to all categories.
• Adding gradient colors to SKILL and Job-Category

patterns = df.Category.unique()
for a in patterns:
    ruler.add_patterns([{"label": "Job-Category", "pattern": a}])

# options=[{"ents": "Job-Category", "colors": "#ff3232"},{"ents": "SKILL", "colors": "#56c426"}]
colors = {
    "Job-Category": "linear-gradient(90deg, #aa9cfc, #fc9ce7)",
    "SKILL": "linear-gradient(90deg, #9BE15D, #00E3AE)",
    "ORG": "#ffd966",
    "PERSON": "#e06666",
    "GPE": "#9fc5e8",
    "DATE": "#c27ba0",
    "ORDINAL": "#674ea7",
    "PRODUCT": "#f9cb9c",
}
options = {
    "ents": [
        "Job-Category",
        "SKILL",
        "ORG",
        "PERSON",
        "GPE",
        "DATE",
        "ORDINAL",
        "PRODUCT",
    ],
    "colors": colors,
}
sent = nlp(data["Resume_str"].iloc[5])
displacy.render(sent, style="ent", jupyter=True, options=options)

SENIOR PROJECT MANAGER Professional Summary Ambitious Construction Executive experienced in commercial construction with over 30+ years DATE of experience. Proactive, resourceful and hardworking with strong follow-through. Excellent problem-solving and time management abilities. Skills Advanced problem solving Project planning and development Finance SKILL and accounting SKILL Employee relations Team building  Negotiations expert Strategic planning Contract review and recommendations Work History Company Name Senior Project Manager | City , State | August 2017 DATE - Current Opening of Denver Operations for McCauley Constructors Establishing ORG protocols, procedures and reporting mechanisms for a satellite location Procurement ORG and management of teams to establish a stand-alone profit center Act as the liaison with main company office representatives to provide information on activities of the satellite location Actively project manage 3 to 5 projects in the $5 to­ $10 million range. Monitored the market to capitalize on the latest trends. Supervised the work of team members, offering constructive feedback on their work performance. Monitored timelines and flagged potential issues to be addressed. Collaborated with the Pre-Construction department ORG to ensure accurate and complete project budgets. Company Name President | City , State | November 2010 DATE - August 2017 DATE Qualified competitive subcontractor bids prior to execution of contracts. Facilitated processing of RFI ORG 's, submittals and samples among the general contractor, the owner and the owner's consultants. Educated general contractor personnel on the quality standards throughout the construction process. Obtained notices of completion and compliance certifications from all of the construction administration consultants. Reviewed and investigated Proposed Change Order Requests ( PCOR ORG ). Stayed consistent with project schedules and plans for all FFE ORG installations. Submitted all project closeout documents in accordance with the contract. Assigned projects and tasks to employees based on their competencies and specialties. Accurately provided status information on project progress to the project management SKILL . Led and managed resolution of all issues during project construction and commissioning phases. Led the planning, budgeting and direction of all construction projects. Carefully coordinated plans and specs using marketing SKILL programming standards. Company Name Senior Project Manager | City , State | January 2006 DATE - November 2010 DATE Managed teams of on-site subcontractors on multiple sites simultaneously. Acted as the liaison with company safety representatives to promote awareness and understanding of safety protocols. Proficiently used the Incident and Issues Tracking ORG ( IIT ORG ) system to document all on site issues. Directed all phases of commercial construction projects, from budgeting  to closeout. Collaborated with the Accounting department ORG to implement electronic accounts payable system. Company Name President | City , State | January 2003 DATE - December 2005 DATE Oversaw business SKILL -wide changes to modernize procedures and organization. Developed program to promote new managers from within, leading to a cohesive leadership structure. Obtained building and specialty permits from local jurisdictional agencies. Conducted weekly DATE production and operations contractor meetings, which facilitated stronger communication and the ability to resolve critical issues. Performed regular job site observations to provide direction for all general contractor personnel. Trained and promoted continued education for all onsite crew members.  Company Name Project Manager | City , State | April 1999 DATE - December 2002 DATE Kept meticulous records of all costs and expenses and analyzed that data against the budget. Managed between 10 to 20 projects per year. Directed changes to the project scope and cost and implemented appropriate change management processes to keep the project on track. Provided outstanding service to clients to not only maintain but to extend the relationship for future business SKILL opportunities. Conducted meetings with clients to determine project intent, requirements and budgets. Company Name Project Manager | City , State | April 1997 DATE - March 1999 DATE Conducted meetings with clients to determine project intent, requirements and budgets. Maintained project schedules by managing timelines and making proactive adjustments. Directed changes to the project scope and cost and implemented appropriate change management processes to keep the project on track. Performed regular job site observations to provide direction for all general contractor personnel. Conducted all critical pre-installation conferences with subcontractors, consultants and manufacturer's representatives. Education Masters of Real Estate & Construction Management Construction Management University of Denver ORG City , State | 1995 DATE Construction Management BBA ORG Business SKILL Management East Tennessee State University ORG City , State | 1991 DATE Business Management Skills Safety Oversight, Prime Contract Negotiations, Budgeting, Competitive Contracts Management ORG ,   Marketing SKILL , Microsoft Office ORG , Microsoft Project ORG , Pro Core Project Management ORG , Personnel Management, Procurement, Project Management SKILL , Systems Protocols and  implementation and Quality Control ORG . Certifications LEED AP ORG ICC Class A License  ASHE ( American Society of Health Engineers ORG )

Your Resume Anlaysis

input_resume="""
Dristanta Das,

Data Science,
Data Analytics,

 
Languages,
Python, 
R,
Systems,
Docker, 
CVAT, MySQL, Neo4J
,Python Modules,
Numpy ,
Pandas,
Matplotlib,
TorchAudio ,
Pytorch ,
Scikit-learn,
Machine Learning,
Regression ,
Classification,
Cross‐validation, Naive Bayes,
Dimensionality Reduction, K-Means,
Decision Tree, Boosting,
Deep Learning,
Multilayer Perceptron, CNN, RNN,
LSTM, Transfer Learning,
Experience
Aug,21’– • Summer Project Intern IITKGP
Under the guidance of professor C.S Kumar ,Mechanical Engineering
Project
• Machine Learning
Feb–June,21’ • 2.5D Visual Sound 
Implemented a deep learning model to convert common monaural
audio into binaural audio by leveraging video.
• Computer Vision
Feb,21’ • Hybrid Image Formation 
Hybrid images are images with two interpretation. The interpretation changes with viewing distance.
Mar,21’ • Harris Corner Detection and SIFT 
The Harris Corner detector is a standard technique for locating interest points on an image.
Apr,21’ • Hough Line and Circle Detector 
May,21’ • Fundamental Matrix using RANSAC 
The objective of this project was to improve upon image matching by
leveraging epipolar geometry of a stereo image pair while applying
RANSAC to reject poorly matched feature points.
• Exploratory Data Analysis
Sept–Dec,20’ • Zomato Restaurants Data 
Analyzed the best restaurants of the cities all around India.
Nov,20’–
Dec,20’
• Player Data of FIFA21 Game 
Analyzed the best players and their performance stats.
• Statistical Data Analysis
Mar–Apr,21’ • Bankruptcy Data 
Analyzed the Bankruptcy data and drew inference out of it. Used
LDA to discriminate the bankrupt and non-bankrupt firms.
Education
2020-Present M.Sc. in Big Data Analytics RKMVERI
GPA: 7.67
Courses
• Machine Learning • Optimisation Algorithm
• Multivariate Statistics • Computer Vision
• Optimisation for Machine Learning • Linear Algebra
• Probability and Stochastic Process • Data Structure and Algorithm
• Artificial Intelligence • Time Series and Survival Analysis
• Online Learning • Data Mining
2017–2020 B.Sc. in Mathematics Presidency University
CGPA: 7.05
2015–2017 Higher Secondary Jodhpur Park Boys’ School
Marks: 84.2%
2005–2015 Secondary Jodhpur Park Boys’ School"""

sent2 = nlp(input_resume)
displacy.render(sent2, style="ent", jupyter=True, options=options)

Dristanta Das ORG ,

Data Science SKILL ,
Data Analytics SKILL ,


Languages SKILL ,
Python SKILL ,
R SKILL ,
Systems,
Docker PERSON ,
CVAT, MySQL GPE , Neo4J SKILL
, Python Modules ORG ,
Numpy PERSON ,
Pandas SKILL ,
Matplotlib,
TorchAudio ,
Pytorch SKILL ,
Scikit-learn,
Machine Learning ORG ,
Regression ,
Classification,
Cross‐validation, Naive Bayes ORG ,
Dimensionality Reduction SKILL , K-Means ORG ,
Decision Tree SKILL , Boosting,
Deep Learning ORG ,
Multilayer Perceptron ORG , CNN ORG , RNN ORG ,
LSTM ORG , Transfer Learning ORG ,
Experience
Aug,21’– • Summer Project Intern IITKGP
Under the guidance of professor C.S Kumar PERSON , Mechanical Engineering SKILL
Project
• Machine Learning SKILL
Feb–June,21’ • 2.5D Visual Sound  Implemented a deep learning SKILL model to convert common monaural
audio into binaural audio by leveraging video.
• Computer Vision SKILL
Feb,21’ • Hybrid Image Formation  Hybrid images are images with two interpretation. The interpretation changes with viewing distance.
Mar,21’ • Harris Corner Detection and SIFT 
The Harris Corner detector is a standard technique for locating interest points on an image.
Apr,21’ • Hough Line and Circle Detector ORG 
May,21’ • Fundamental Matrix using RANSAC ORG 
The objective of this project was to improve upon image matching by
leveraging epipolar geometry of a stereo image pair while applying
RANSAC ORG to reject poorly matched feature points.
• Exploratory Data Analysis SKILL
Sept–Dec,20’ • Zomato Restaurants Data 
Analyzed the best restaurants of the cities all around India GPE .
Nov,20’–
Dec,20’
• Player Data of FIFA21 Game ORG 
Analyzed the best players and their performance stats.
• Statistical Data Analysis SKILL
Mar–Apr,21’ • Bankruptcy Data 
Analyzed the Bankruptcy ORG data and drew inference SKILL out of it. Used
LDA to discriminate the bankrupt and non-bankrupt firms.
Education
2020-Present M.Sc. in Big Data Analytics ORG RKMVERI
GPA ORG : 7.67
Courses
• Machine Learning SKILL • Optimisation Algorithm SKILL
• Multivariate Statistics SKILL • Computer Vision SKILL
• Optimisation for Machine Learning SKILL • Linear Algebra
• Probability and Stochastic Process • Data Structure SKILL and Algorithm SKILL
• Artificial Intelligence SKILL • Time Series SKILL and Survival Analysis
• Online Learning • Data Mining SKILL
2017–2020 B.Sc PRODUCT . in Mathematics Presidency University ORG
CGPA: 7.05
2015–2017 Higher Secondary Jodhpur Park Boys’ School Marks ORG : 84.2%
2005–2015 Secondary Jodhpur Park Boys’ School

Match Score

In this section, I am allowing recruiters to add skills and get a percentage of match skills. This can help them filter out hundreds of Resumes with just one button.

input_skills="Data Science,Data Analysis,Database,MySQL,Machine Learning,Deep Learning,Analytics,Artificial Intelligence,python,pytorch"

req_skills = input_skills.lower().split(",")
resume_skills = unique_skills(get_skills(input_resume.lower()))
score = 0
for x in req_skills:
    if x in resume_skills:
        score += 1
req_skills_len = len(req_skills)
match = round(score / req_skills_len * 100, 1)

print(f"The current Resume is {match}% matched to your requirements")

The current Resume is 80.0% matched to your requirements

print(resume_skills)

['python', 'decision tree', 'machine learning', 'dimensionality reduction', 'artificial intelligence', 'multivariate statistics', 'exploratory data analysis', 'pandas', 'docker', 'inference', 'big data', 'computer vision', 'data science', 'mechanical engineering', 'languages', 'time series', 'algorithm', 'analytics', 'data analysis', 'pytorch', 'deep learning', 'data structure', 'data mining', 'corner detection']

Topic Modeling - LDA

LDA, or Latent Dirchlet Allocation is arguably the most famous topic modeling algorithm out there. Out here we create a simple topic model with 4 topics. The code was inspired by Allan's project: Topic Modeling of NLP GitHub repositories

docs = data["Clean_Resume"].values
dictionary = corpora.Dictionary(d.split() for d in docs)
bow = [dictionary.doc2bow(d.split()) for d in docs]
lda = gensim.models.ldamodel.LdaModel
num_topics = 4
ldamodel = lda(
    bow, 
    num_topics=num_topics, 
    id2word=dictionary, 
    passes=50, 
    minimum_probability=0
)
ldamodel.print_topics(num_topics=num_topics)

[(0,
  '0.012*"state" + 0.011*"company" + 0.010*"city" + 0.009*"management" + 0.008*"name" + 0.007*"business" + 0.005*"client" + 0.005*"financial" + 0.005*"skill" + 0.005*"marketing"'),
 (1,
  '0.008*"engineering" + 0.008*"project" + 0.008*"state" + 0.007*"system" + 0.007*"equipment" + 0.007*"city" + 0.006*"company" + 0.005*"training" + 0.005*"management" + 0.005*"name"'),
 (2,
  '0.021*"customer" + 0.013*"state" + 0.012*"city" + 0.012*"sale" + 0.011*"company" + 0.011*"service" + 0.009*"name" + 0.007*"skill" + 0.007*"product" + 0.007*"food"'),
 (3,
  '0.013*"project" + 0.012*"system" + 0.009*"management" + 0.007*"team" + 0.007*"design" + 0.007*"support" + 0.007*"software" + 0.007*"company" + 0.007*"network" + 0.006*"application"')]

pyLDAvis

The best way to visualize Topics is to use pyLDAvis from GENSIM.

• topic #1 appears to relate to the project,management,system.
• topic #2 relates to management,company,business.
• topic #3 relates to customer,services , state.
• topic #4 relates to state,city,student.

pyLDAvis.enable_notebook()
pyLDAvis.gensim_models.prepare(ldamodel, bow, dictionary)

/usr/lib/python3/dist-packages/past/builtins/misc.py:45: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  from imp import reload
/usr/lib/python3/dist-packages/past/builtins/misc.py:45: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  from imp import reload
/usr/lib/python3/dist-packages/past/builtins/misc.py:45: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  from imp import reload
/usr/lib/python3/dist-packages/past/builtins/misc.py:45: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  from imp import reload
/usr/lib/python3/dist-packages/past/builtins/misc.py:45: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  from imp import reload
/usr/lib/python3/dist-packages/past/builtins/misc.py:45: DeprecationWarning: the imp module is deprecated in favour of importlib; see the module's documentation for alternative uses
  from imp import reload